Inventory how-to

How to measure forecast accuracy

Measure forecast accuracy by comparing what you predicted you would sell against what you actually sold, per variant, over a fixed period.

Short answer: Measure forecast accuracy by comparing what you predicted you would sell against what you actually sold, per variant, over a fixed period. The simplest useful method is Mean Absolute Percentage Error (MAPE): take the difference between forecast and actual, divide by actual, and average that across your SKUs.

Why measure it at all

If you never check forecast accuracy, you cannot tell if you are ordering too much or too little for the wrong reasons. A bad forecast shows up later as dead stock or stockouts. Checking it regularly lets you catch problems early and adjust safety days or lead times instead of guessing.

Pick a simple error measure

MAPE (Mean Absolute Percentage Error) is the easiest to explain to anyone on your team. For each SKU: error = forecast minus actual. Take the absolute value, divide by actual sales, multiply by 100 to get a percentage. Average this across SKUs or across weeks to get an overall number.

Another option is Mean Absolute Error (MAE), which just averages the absolute unit difference without turning it into a percentage. MAE is useful when you want to know "how many units off" rather than "how far off in percent," especially for low-volume SKUs where percentages swing wildly.

Measure at the right level and frequency

Check accuracy per variant, not just per product or store-wide. A store-wide average can hide the fact that your best seller is badly forecast while slow movers look fine. Weekly or monthly reviews work for most small stores. Daily checks are usually too noisy to act on.

Watch for patterns, not single misses

One bad week does not mean the forecast method is broken. Look for a SKU that is consistently over or under forecast for three or more periods in a row. That is a signal to adjust manually, for example after a promotion ends, a supplier changes lead time, or a product starts a new season.

Step by step

  1. List forecast quantity and actual sales for each SKU at the end of each period
  2. Calculate the absolute difference between forecast and actual for each SKU
  3. Divide that difference by actual sales and multiply by 100 to get percent error (MAPE)
  4. Average the percent errors across your top SKUs to get a single accuracy score
  5. Flag any SKU with consistent errors over three or more periods
  6. Adjust safety days or your own manual override for those flagged SKUs
  7. Re-check accuracy the next period to confirm the adjustment worked

Worked example

Say your forecast for a candle SKU was 100 units for the month, and actual sales came in at 80 units. The error is 100 minus 80 = 20 units. Divide 20 by actual sales of 80, giving 0.25, so the percent error is 25%. If a second SKU had a forecast of 50 and actual of 55, the error is 5, divided by 55 is about 9%. Averaging 25% and 9% gives a rough MAPE of 17% across these two SKUs for the period.

How SteadyShelf helps

SteadyShelf does not currently show a built-in forecast accuracy report or MAPE score. You can still do this manually by exporting the suggested order quantities and your actual sales from your store admin, then running the calculation above in a spreadsheet each month.

$29/month flat, unlimited products and suppliers, 30-day free trial. Run the numbers now with the free reorder point calculator.

SteadyShelf is launching on the Shopify App Store. Join early access to get the install link first.

Related guides

All inventory how-to guides ยท Inventory glossary

FAQ

What counts as a "good" MAPE score for a small store?

There is no universal good number since it depends heavily on SKU volume and variability. Low-volume or highly seasonal SKUs will naturally show higher percent errors than steady best sellers, so judge each SKU against its own history rather than a fixed target.

Should I measure accuracy on units or on revenue?

Units are usually more useful for inventory decisions since you order and receive stock in units, not dollars. Revenue-based accuracy can be worth tracking separately for finance purposes, but it can mask unit-level stockout or overstock problems on lower-priced items.

Let SteadyShelf do this math for every SKU

Order-by dates per supplier, seasonal forecasts and draft POs for your Shopify store. $29/month after a 30-day free trial.

Get early access